It was measuring the text to try to determine the size that it needed to
scale the emoji character to. Unfortunately it was accidentally trying
to measure the character under the cursor instead of the emoji character
itself, which is wrong, but more seriously doesn't work at all when the
cursor is at the end of the line.
This was already fixed before in change 144474, but that change never got
merged over to donut. So this merges it now.
Bug 2087915
Change-Id: Ib4804d330a029a966207b3b07271f84e6b2652c0
- Reduce the amount that we ask processes to GC after a significant
operation occurs, but introducing a minimum time between GCs and
using this in various ways to schedule them.
- Don't spam all of the processes with onLowMemory(). Now deliver
these using the same gc facility, so we do the processes one at a
time, and don't allow the same process to get this call more than
once a minute.
- Increase the time a service must run before we will reset its
restart delay to 30 minutes (from 10).
- Increase the restart delay multiplication factor from 2 to 4.
- Ensure that we don't restart more than one service every 10 seconds
(unless some external event causes a service's process to be started
for some other reason of course).
- Increase the amount of time that a service must run before we
decide to lower it to a background process.
And some other things:
- Catch IllegalArgumentException in ViewRoot like we do for no
resources to avoid the system process crashing.
- Fix a number of places where we were missing breaks between the
activity manager's message dispatch func(!!).
- Fix reason printed for processes in the background.
- Print the list of processing waiting to GC.
Before, after using the Browser, memory-hungry apps could
become very sluggish. This was because the search dialog in the
system process had the BrowserProvider open, which in turn had
EnhancedGoogleSearch open. Since EhancedGoogleSearch runs in acore,
the system would keep both the Browser process and acore to stay
around forever.
The cause (or at least one common cause) for this was that
if the user types quickly, and clicks on a suggestion before
the displayed suggestions have caught up, some suggestion cursors
are not be closed.
This change solves this problem by adding a close() method to
SuggestionsAdapter. SuggestionsAdapter now closes any cursors
that are passed to it after close() is called.
Fixes http://b/issue?id=2078226
"global search holding reference to browser: system -> browser -> acore = :("
- @hides relevant APIs
- removes relevant javadoc
- enforces that only system apps can participate
note: general support is still there, will be easy to reenable when we are ready.
* changes:
DOCS ONLY. add manifest documentation for uses-feature and supports-screens elements. also update the navigation and manifest home page, update the uses-sdk element to include new maxSdk and targeSdk attributes, and add some sample code to DisplayMetrics to query the device for screen info.
add manifest documentation for uses-feature and supports-screens elements.
also update the navigation and manifest home page,
update the uses-sdk element to include new maxSdk and targeSdk attributes,
and add some sample code to DisplayMetrics to query the device for screen info.
* changes:
Fix bug 2043140. A race condition is encountered when an application invokes shutdown() on its TextToSpeech object while is has speak() requests still running. Since the TTS service destructor releases the synthesizer resources and sets the corresponding synth reference to null, an NPE was observed. The fix consists in catching NPEs whenever the sNativeSynth object is accessed, and return the matching error for the call. This change is a "low risk" version of the fix for bug 2025765i (same issue) which was reverted because it was higher risk than this CL: it affected the logic of each call to sNativeSynth. This CL only sets an error code when an NPE is fired because sNativeSynth is null.
A race condition is encountered when an application invokes shutdown()
on its TextToSpeech object while is has speak() requests still running.
Since the TTS service destructor releases the synthesizer resources and
sets the corresponding synth reference to null, an NPE was observed.
The fix consists in catching NPEs whenever the sNativeSynth object is
accessed, and return the matching error for the call.
This change is a "low risk" version of the fix for bug 2025765i (same
issue) which was reverted because it was higher risk than this CL:
it affected the logic of each call to sNativeSynth. This CL only sets
an error code when an NPE is fired because sNativeSynth is null.
Includes:
- an overview section on including search suggestions in quick search box
- additional table of searchable metadata attributes related to quick search box
- additional column descriptions that we've added for quick search box.
- description of how to expose your content provider's search suggestions to
the search infrastructure if you have permissions guarding your content provider.
* changes:
First update of SearchManager docs to reflect the new Donut Quick Search Box feature. This simply cleans up any inconsistencies or weirdnesses, we still need to add the new info for QSB.
When RemoteViews inflates a layout and set of actions, it
uses a cloned remote Context so it can fully access the
resources needed to complete the inflate. However, in one
specific case when inflating the root view, LayoutInflater
calls back to the soon-to-be parent view to request the
correct set of LayoutParams through generateLayoutParams().
The default implementation of generateLayoutParams() uses
the Context of the /local/ View instead of the remote
Context being used for the inflation, which means that any
LayoutParam references in the root view are unresolved.
To fix this, we internally create our own copy of the target
remote Context. This way, when the remote LayoutInflater
calls generateLayoutParams(), we inflate it using the remote
Context to correctly resolve any references.
This change is required as part of http://b/2038594
close itself directly because it may not happen correctly for some cursors
currently. This fixes http://b/2036290, which is being caused by
http://b/2015069 which we are not fixing for Donut, so this is a hack around
that for the time being.
* changes:
Close the search dialog before starting voice search in all cases, i.e., also when launching web search, not just when it's using the intent API.
* changes:
Fix#2025900. If a https request is canceled while openConnection is called and a ssl error happened, the http thread can be blocked for up to 10 min. Added code to detect this case and unlock the thread.
search. Without this, because of the recent changes to how the search
dialog hides and resumes, we end up re-showing the search dialog after
returning to the app from voice search, obscuring the new voice search
results behind.
Fixes http://b/2025364
When the suggstion provider is in a different package from the searchable
activity, the icons stored for shortcuts used the package of the
activity instead of the provider.
Fixes http://b/issue?id=2023993
The drawing matrix used in onDraw() is mDrawMatrix, but in setImageMatrix()
only mMatrix is updated. We need to call configureBounds() to update
mDrawMatrix if it was null. The is found in bug 2023281.
